Rob Landers
Let’s face it, we’re all pretty terrible at estimating things. Erik Bernhardsson takes the data and runs with it to generate a statistical model. It would be fun to use p2 (our internal team blogs) to come up with our own distribution. Maybe in a future post!
Carly Stambaugh
I love video games. (I’m currently on my 3rd play through of Breath of the Wild.) I read an article about AI upscaling, a process that transforms low resolution game graphics to high resolution, making them look more modern. Using a GAN trained on high and low resolution versions of the same photo, AI upscaling allows for much better results than commonly used nearest neighbor algorithms. With AI upscaling tools more widely available, many classic games are getting a new look. But, it’s no quick fix. Even with the aid of these tools, it still takes dozens of hours to upscale a game. I guess it’ll be a while before my favorite Play Station 1 game, Skullmonkeys, looks as good as BOTW.
Boris Gorelik
Moving to a World Beyond “p < 0.05” is an opinion article written by three scientists that call for abandoning significance tests in science. This same problem was also featured on nature.com.
Data for Breakfast 
Leave a comment